June 12, 2019
One of our goals at ATOM is to optimize preclinical safety predictions in silico, so we can incorporate predictive toxicology early in the drug discovery process. Working toward this challenging goal is of utmost importance in the pharmaceutical industry as well as in academic drug discovery as its success will not only save the pharmaceutical industry billions of dollars, but also accelerate the drug discovery process. A drug discovery platform that can reliably predict safety will result in a reduced rate of clinical trial failures, leading to a higher proportion of new molecular entities (NMEs) receiving regulatory approval and reaching patients in need. ATOM is contributing to this goal by developing and profiling new cell-based and complex in vitro assays, collecting and generating world-class datasets, and integrating those into advanced computational models that can better predict drug safety.
According to the US Food and Drug Administration (FDA), drug induced liver injury (DILI) is the most common cause of safety-related drug marketing withdrawals [1]; it is also a leading cause of attrition during drug development. While widely recognized as a major challenge, a consensus on the ideal set of predictive in vitro and in silico tools to detect and investigate DILI has yet been reached. In an attempt to address this need, we are building an experimental and computational platform with which we can reliably measure and predict DILI at the organism and population level. Within the overarching ATOM drug discovery platform, this safety module will then be incorporated into ATOM’s generative molecular design loop to help generate new and optimized candidate molecules that are not only efficacious, but also safe.
In our quest to build a robust predictive DILI platform, we have challenged ourselves to answer a question that remains elusive: What is the minimal set of predictive in vitro tools to detect and investigate DILI? Evidence suggests that data from bile acid (BA) transporter inhibition assays are a good predictor of BA-mediated DILI [2, 3]. Recent work has also combined cytotoxicity with additional end points, including inhibition of hepatic transporters, and evaluation of mitochondrial dysfunction for a comprehensive assessment of DILI risk [4]. Implementing assays that can assess subtle sub-lethal markers is thought to be of high priority as they can better account for the hepatic effects observed in vivo [4].
With this in mind, we have launched into an exciting and challenging experimental endeavor to develop and profile multiple assays. These include assays with mechanistic end points, such as inhibition of hepatic transporters, and multiple high-content assay formats, to measure and predict hepatocyte cell health using not only markers for cytotoxicity, but also more subtle sub-lethal markers including markers that measure mitochondrial membrane potential, reactive oxygen species and glutathione content. We are also coupling these data with live measurements of the oxygen consumption rate of hepatic cells. Together, these multi-parametric and mechanistic data, coupled with quantitative systems toxicology tools and machine learning, will enable us to predict DILI from the structure of a proposed drug lead.
Furthermore, there is also a need to evaluate more relevant in vitro models that potentially better mimic the in vivo response in humans, including long term exposure and inflammatory components. We aim to further develop our safety platform to address this unmet need by systematically assessing the predictive accuracy of not only the 2D cell-based models, but also more complex in vitro models such as organoids and organs-on-chips.
Sarine Markossian, PhD
Arkin Lab, Department of Pharmaceutical Chemistry, School of Pharmacy, University of California San Francisco
References:
1. FDA. Guidance for Industry. Drug-Induced Liver Injury: Premarketing Clinical Evaluation. 2009.; Available from: https://www.fda.gov/media/116737/download.
2. Morgan, R.E., et al., A multifactorial approach to hepatobiliary transporter assessment enables improved therapeutic compound development. Toxicol Sci, 2013. 136(1): p. 216-41.
3. Pedersen, J.M., et al., Early identification of clinically relevant drug interactions with the human bile salt export pump (BSEP/ABCB11). Toxicol Sci, 2013. 136(2): p. 328-43.
4. Chen, M. and Y. Will, Drug-induced liver toxicity. Methods in pharmacology and toxicology,. 2018, New York, NY: Humana Press. 667 pages.
